CompStak launches Agent Connect, an MCP server for verified commercial real estate data. September 29, 2026 Agent Connect (MCP) is available with CompStak One Connect, a new tier that also adds Rent Predictor to the CompStak One platform. [New York, NY, September 29, 2026] - CompStak, the leading source of commercial real estate lease comps, sale comps, property data, and analytics, today announced the launch of Agent Connect, an MCP server that lets teams query CompStak's verified data from inside the AI tools they already use, like Claude, ChatGPT, Copilot, and Gemini. Agent Connect is available through CompStak One Connect, a new offering that pairs the CompStak One platform with Agent Connect and Rent Predictor. CRE teams have spent the past few years building internal tools and AI workflows on top of their own data. Bringing a verified external dataset into those workflows required a complex and time-consuming integration project that engineering teams often do not have time to build. Agent Connect runs on Model Context Protocol, an open standard for connecting AI assistants to outside data sources. With it, an analyst can ask for a comp set in the assistant already open on their desktop and get an answer drawn from CompStak's verified database. "I use AI agents every day, and so do a growing share of our customers. CRE work is moving out of standalone platforms and into agents and custom workflows. At CompStak, we're building for the future, not pretending it isn't happening. Agent Connect puts CompStak's verified comps, plus our Market Insights, Property Insights, and Space Insights APIs, inside the tools our customers already have open. An analyst can go from a question to a sourced comp set or a first-draft investment memo in one session, with no integration project and no data export. The end result provides not only a magical experience for users, but a deliverable they can stand behind," said Michael Mandel, Co-Founder and CEO of CompStak. How Agent Connect Works. * Ask in plain language: Teams query comps, aggregations, and market statistics without writing a query or exporting a file. * Works alongside internal data: CompStak results arrive in the same session as the customer's internal models and documents. * Works with existing AI assistants: Agent Connect connects to Claude, ChatGPT, Copilot, Gemini, and internally built agents. CRE teams can now direct their AI agents with tasks like: * Compare average office net effective rents since 2020 in Hudson Yards, FiDi, and Midtown South. * Benchmark my portfolio's lease rollover risk against the LA industrial market overall in 2027. * Price my portfolio's upcoming vacancies. * Build me an investment memo, with complete lease and sale comps, mark to market analysis, and risk analysis. Compare this potential purchase to recent sales and properties in the competitive set. Also new in CompStak One Connect. Rent Predictor produces Estimated Starting Rent Today, a modeled starting rent for spaces with no recent comp, powered by CompStak's proprietary AI and trained on millions of verified lease comps. Most space in a given market has not traded recently, and teams have either estimated it by hand or left it out of the model. Rent Predictor gives them a number for that space. Built on CompStak One. One Connect includes everything in the CompStak One platform: * Verified lease, sale, property, and loan data across 105 US markets * Analytics for market and submarket trends, including net effective rent and lease term * Portfolios for benchmarking assets against custom competitive sets * CompScout for natural-language comp search * AI Market Summary for on-demand market narratives Every record reaching those tools comes through CompStak's verification process: a network of more than 40,000 brokers, appraisers, and researchers submitting comps, machine learning and statistical anomaly detection flagging figures outside expected ranges, and in-house research analysts reviewing transactions and adding detail. Proprietary CRE Ontology and Entity Resolution connect properties, owners, and transactions into a single source of truth. About compstak. CompStak is the leading source of commercial real estate lease comps, sale comps, property data, and analytics. Comps are crowdsourced from a network of more than 40,000 brokers, appraisers, and researchers. Every comp is verified before it is published: first by machine learning and statistical anomaly detection, then by a team of CRE data analysts. Lenders, owners, investors, brokers, and appraisers use that data to underwrite and value assets across 105 US markets, working within the CompStak platform, in their own models, or in the AI tools they already use. To learn more about CompStak, visit compstak.com. Media inquiries: Brigette Palombo

Columbia CompStak (CCRI) Rent Index national update: May 15, 2026. May 15, 2026 Commercial real estate rent data is only as useful as the methodology behind it. That's why CompStak partnered with Columbia Business School faculty to build the Columbia CompStak Rent Index (CCRI), the first quality-adjusted CRE rent index tracking constant-quality net effective rents across U.S. office, retail, and industrial markets, both nationally and across major MSAs. Unlike traditional rent benchmarks that track asking rents or simple averages, the CCRI controls for quality mix and incorporates concessions, giving investors, lenders, occupiers, and capital allocators a cleaner read on where rents are actually moving. Each month, CompStak publishes a national update powered by its database of verified lease transactions, one of the largest and most comprehensive in commercial real estate. Below is a breakdown of what the latest data shows across all three property sectors at the national level, as of May 15, 2026. Constant-Quality net effective rent indices. Nationally, office effective rent growth has been rebounding over the past six months. National retail rents are down in March but roughly flat over the past year. Industrial rents are down over the past six months. The relative performance of office and industrial is a reversal of fortunes compared to the prior five-year period. Most recent observation: March 2026. Manhattan Office NER growth is up 8.7% over the year to 2025.Q4. NER rent growth was down 1.17% in 2025.Q4. Its next update on June 15, 2026 will include the first estimates for 2026 Q1. Preliminary analysis shows Manhattan office rent growth will be strong. Office: Charlotte (+31.6%), Baltimore (+22.2%), and Austin (+18.5%) lead office rent growth over the year to 2025.Q4, while coastal markets show divergence. The New York MSA posts healthy gains (+12.1%) but San Francisco (−5.2%), Seattle (−6.5%), and Miami (−8.2%) continue to struggle. The national Office index registered a +4.54% over the same period, with strong performers offsetting weakness in coastal MSAs. Retail: Atlanta stands out with +84.1% rent growth on relatively thin volume, accompanied by other top performers like Detroit (+36.4%) and Los Angeles (+23.2%). Houston (−20.1%), Boston (−8.8%), and Las Vegas (−7.4%) show the biggest declines. The national Retail index registered +4.71% NER growth over the same period. Industrial: Industrial growth is more compressed across markets compared to Office and Retail, with Boston leading with +11% rent growth and Charlotte declining the most at −17.6%. The national Industrial index posted −0.75% over the year to 2025.Q4, showing that, in aggregate, the industrial real estate boom is over.